Today we are actually gonna get into artificial intelligence, so Azure has machine learning and it is connected to NAV or can be connected to NAV in a simple form, however, it's not very intuitive hot to connect it so it's a little bit puzzling.
If I go here into items for example, and I’m here on the ATHENS mobile pedestal, I can go on the sidebar here or the boxes and I have something called forecast and it comes up with not enough historical data, actually when you log into CRONUS or you setup a new company its always says this, I think this is misleading, I don't know where this is coming from, I dont think it's coming from anywhere I just think its say not enough historical data always, but in fact it's not connected.
Let's go ahead and look at the setup here. If I get into the setup of this, its setup with some standard parameter and there is an API a URI and API key with here, this so here nothing is set up, which means it's not connected to an artificial intelligence or machine learning, but the forecast is supposed to be.
Let's go ahead and connect it an see if we can get something going. So if I click here on open Cortana intelligence model, I actually get into kind of a splash screen here, which is forecasting for dynamics, and it tells you here an example API, an example endpoint URI, you cannot use this, this is just an example, it won't work.
You actually have to open up the artificial intelligence studio. So if you get into that, if you are actually setup an account here, it's very easy to set that up, I have one already, it will open up this experiment inside the artificial intelligence so don't be too concerned with all of this, this is of course very heavy right here in technical stuff and this is actually the experiment or they call it experiment. But what it is really a forecasting model that has been created for NAV inside the Azure machine learning studio. This hangs out in the cloud and when NAV need to forecast something it actually connects to this, gets the forecasting, brings it back into NAV and you see it there on the fact box. So this is the programming behind and you can actually change that.
Now if I just wanna get this going in its basic form, what I need to do is actually first save this, so I’m gonna save this experiment right here, so it's saved in my workspace and then I have to actually run it, so when I’m running it, the system goes through everything and you can see queues up, the experiment, and then it starts going into you know the sample data actually get the R script, slips the data. And then it completes.
So, once it ran, you just run it, let it go through its motions, that is, you can actually say deploy web service and what that does, is form your work space in Azure, it takes this model and it deploys it as a web service, and so here I finally get my API key , this is the API key I can use. So I just take that, copy it here and go back into NAV right here and I’m going o paste that into the key, now I also need a URI.
Where do I get that?, If I click here into the request-response page, that sort of a splash page that allows you to read about how to connect this remotely. The actual URI is gonna be here, so you can copy . It should be much easier. But the good part about this is you can basically create your own forecasting module inside the artificial intelligence and us that. But anyways.
I set this up right now and now I wanna do update forecast. Now it says: sales forecasts are being updated in the background and it takes a few moments before they are available. Ok, go out here, let me just go ahead and take a look at here, Paris guest chair, and now I can update it straight here, now I can see I get a forecast for the Paris guest chair, and I’m just gonna take a look at the entries really quick.
You can see the forecast comes up here. So just quickly on the entries I have created some sales with month intervals: 10, 14, 12, 15, 18 so you guys can guess the neighborhood that I’m here, the idea was that I get somewhere between 10 and 18 I guess suggestions, if this all worked out fine.
If I close that out and take a look at this, I get here, 1,200. Ok so don't be alarmed, the algorithm actually worked, this is the inventory balance, I already had 1,230 on hand and so right now the inventory balance is going down. If I only want to see my sales forecast I have to click here, and now I see it. So you can see that is hanging around 15, Actually just give me straight 15 all the way through so that's a pretty boring forecast but at least it's doing its job.
From here I can actually create a purchase order to buy or purchase invoice to buy, but now it is connected and its given me data on this. If I go into the Athens pedestal and update that one, I get no historical data so, I think you at least have to have maybe three or four periods to get some data and I had that on the Paris chair guest here.
What would be nice is actually be able to connect this to the real forecast in the NAV, the inventory forecast, that would be great. I’m gonna take a look at that in a little bit and see how we can work that out, maybe we can, maybe we cant. Until next video, if you got something out of this, at least you got the forecast going.
You've learned some of the many ways your business can benefit from using iNECTA Food. It's trusted by other businesses in your industry and can work just as well (or better) for you. iNECTA Food is easy to implement. It's easy on your budget with monthly subscription payments. Take the next step to ensure your food business is running the right kind of business software. Software for your food safety, increased efficiency, and stronger revenue growth.
Curious about implementation costs, training, getting a customized demo & detailed estimate, booking a discovery call?